Considerable effort has been expended in attempts to ameliorate or prevent adolescent alcohol abuse. Unfortunately, such efforts have met with little success. The importance of the problem is clear: not only does teenage problem drinking entail significant societal costs, but moreover the teenager years can be understood as a "window of vulnerability" during which alcohol is first used and the ground work for later abstract. The very limited success in this area highlights the need to further our understanding of the parameters of teenage drinking. With this aim in mind, the original grant initiated a longitudinal investigation into adolescents' cognitive expectancies for the reinforcing effects of alcohol. Based on an extensive literature documenting expectancies' importance in determining drinking consequences, the original study produced evidence of expectancies' powerful role in the longitudinal prediction of early adolescent drinking onset. Specifically, the proposed continuation investigation will capitalize on the existing longitudinal program in the following ways: 1. extend the collection of longitudinal data into the junior and senior years of high school. The ability of adolescent expectancies to sustain their predictive power into these years, when quantity and frequency of adolescent drinking increase dramatically, will be assessed through the collection of expectancy, alcohol consumption, and related variables. This extends the longitudinal period to five years. 2. Further investigate expectancies' diagnostic usefulness. Data will have been obtained from a large sample of adolescents throughout their junior high and high school years. Students will be classified according to drinking style and drinking onset, and cutting score procedures will be applied to expectancy scores to examine expectancies' performance in predictively classifying subjects. 3.Conduct a pilot study of a prevention strategy based on the findings of the original grant. Specifically, an attempt will be made to modify the expectancy scale most predictive of adolescent drinking. This will provide information critical to the development of a more extensive prevention program. 4. Use of this ongoing longitudinal program provides an efficient way to obtain valuable data covering five critical, transitional years in adolescent drinking.